New tool unites human insight with ‘best features of AI’

technology_aiBrands are being offered the chance to develop highly tailored AI-driven marketing campaigns with the launch of a new tool that combines human insight with the best elements of “supervised AI”, giving them more control and solving the problem of poor targeting and wasted spend.

Vignette has been developed by London-based data consultancy Metrix Data Science, and harnesses the power of supervised AI algorithms to tailor content to individual consumers, promising substantial improvements to campaign results.

The supervised system combines AI software and human expertise. This means that, although there could potentially be hundreds of thousands of versions of campaign content, there will be no nasty surprises, because the approach avoids the danger of an errant algorithm serving inappropriate content.

Metrix Data Science argues that in an unsupervised solution there could be more content combinations than stars in the universewith marketers having no clear idea which consumers will see.

Initially, Vignette is being deployed for email marketing campaigns. Email remains one of the most commonly used customer acquisition channels, with huge campaign volumes that would benefit from the use of AI. However, many organisations’ marketers are unsure how to get the best from AI, which can lead to problems with delivery and poor results.

Vignette firstly considers previous email campaign results, such as delivery time of day, subject lines and creative to understand what will work best for each target consumer.

Next, the system builds a Vignette profile of individual consumers, based on factors such as Amazon Prime membership, attitudes to online advertising, demographic information, hobbies, health and financial wellbeing, preferred supermarket, proximity to stores and facilities, and many other variables.

Using machine learning classification algorithms, the system then identifies a pool of data to match the brand’s campaign objectives.

The essential elements of an email campaign – subject line, headline, copy and images – are created using Large Language Models. These are combined with the data to produce Vignettes: despite potentially huge numbers of content variations, highly customised templates are tailored to each recipient.

These are initially run through test campaigns before they are rolled out more widely to prospects. It is claimed that initial results have shown that Vignette significantly enhances both email open and click-through rates when deployed for campaigns.

Metrix Data Science founder and chief executive Bill Portlock said: “We believe Vignette cracks the code for marketing and takes the lead on solving the problem of AI being left alone to ride roughshod, resulting in badly targeted big volumes and wasted spend.

“Crucially, the new system uses a unique and powerful mix of ‘part automation’ to give marketers more control: practical human insight while also benefitting from the best features of AI.

“Email remains the key customer acquisition channel but many brands and tech providers don’t do it well. It means Vignette delivers tailored content with AI functionality. It creates and delivers hundreds of thousands of emails that are perfectly personalised to each individual recipient, while marketers can rest safe in the knowledge that the consumer isn’t receiving any inappropriate or irrelevant communications.

“Organisations with big email campaign volumes are already benefitting from a more efficient and effective way to recruit new customers.”

Related stories
DVLA signs up for StrappedUK cost of living data tool
New tool provides insight into UK cost of living crisis
DVLA reappoints Marketing Metrix to data business
Marketing Metrix nets DVLA data segmentation task
Marketing Metrix expands as DVLA renews contract